Page 15 - Stöck_alii_2008

BMC Evolutionary Biology 2008, 8:56 http://www.biomedcentral.com/1471-2148/8/56

was HKY+G (AIC; 10 M generations; burnin = 1500, the stitutional time (τ = 2µt; t is the time in generations, µ the
same settings were used for the 16S rRNA sequences). For substitution rate per locus and per generation [89]).
RAG-1 fragments (860 bp, model: GTR+G; AIC, burnin = Graphically, the mismatch distribution of a recently
1000) and the alpha-tropomyosine intron (612 bp; expanded population is unimodal and smooth; the wave-
model: HKY+G; AIC; because of the occurrence of inser- shaped curve is centered nearer the y-axis the more recent
tions and deletions, all missing and ambiguous data were the expansion, moving away as the number of mis-
excluded and only 405 bp were analyzed; 5 M genera- matches increases [89]. By letting θ become very large, θ 0
1
tions; burnin = 1000). and τ are estimated from the data [90]. We set θ to
1
1,000,000. We assessed the deviation of the observed dis-
Because amplification or alignment of markers was not tribution from the expected under a model of sudden
equally possible for identical outgroup taxa from the expansion by comparing the raggedness statistics of the
available material, more than one species had to serve as observed distribution with a simulated distribution to
outgroup. In previous analyses [[28] and unpublished determine the probability that the raggedness of the
data], we demonstrated that all used taxa (Bufo surdus, B. observed distribution could have arisen by chance. Third,
raddei, B. calamita, B. bufo) represent suitable outgroup we estimated Tajima's D [91] in DnaSP for each grouping.
species. Tajima's test of selective neutrality compares two θ estima-
tors and its significance is evaluated by comparison of the
Taxonomic subdivision test statistic (D) with values randomly generated under
To evaluate the distinctiveness of genetic groups of West- "neutrality." Significant values indicate the population
Mediterranean green toads of the B. viridis subgroup has deviated from neutrality, or that another demographic
[sensu [28]], and of further geographic subdivisions force has caused the deviation from expectation, such as
within each genetic group, we calculated pairwise F val- population expansion (significant negative D).
ST
ues between all groups using Arlequin 2.0 [85] based on
the mitochondrial control region. Divergence times among the main mitochondrial lineages
were estimated using a Bayesian-coalescence approach, as
Demographic analyses and divergence time estimates implemented in BEAST 1.4.6 [92-95]. In analysis of the
We assessed the possibility of population expansion for control region, we used a matrix of 60 individuals and 752
green toad groupings assembled into different geographi- bp. We started the search with an UPGMA tree, constrain-
cal populations. We applied three different tests, each ing the clade B. boulengeri-B. siculus from Sicily to be
with different strengths, to the mitochondrial control monophyletic. Because we were analyzing a species-level
region dataset (846 or 541 bp), to detect evidence of phylogeny, we used a Yule tree prior, which assumes a
recent expansion. First, we applied Fluctuate [86], a maxi- constant speciation rate per lineage. We applied an uncor-
mum-likelihood estimator of the parameters θ and g (θ = related relaxed molecular clock, with the substitution rate
2Ne/µ; g = exponential population growth rate parame- of the branch lengths being sampled from a prior normal
ter). The exponential growth parameter (g) was used to distribution with a mean value of 0.02 and a standard
estimate the size of the population at time in the past from deviation of 0.007 [94]. The search was conducted for a
N = θ e-(gµ)t where N is the effective population size at time range of substitution rates that varied from 1% to 3% per
t
t
t in the past [86]. Using this equation, t was estimated by million years. We ran four independent analyses for 20 ×
6
substituting N with N /N t = 0 = 0.1 (µ is DNA substitution 10 generations. We checked for convergence and station-
t
t
rate per site per generation, N is the female effective pop- arity of the different analyses in Tracer 1.4 and combined
t
ulation size at time t). Repeated analyses to ensure stabil- the results in the BEAST module LogCombiner 1.4.4 (after
ity of estimates were run as described by Stöck et al. [28]. removing the first 2 × 10 generations from each analysis
6
Growth was inferred using logarithmic likelihood ratio as "burnin").
tests with one degree of freedom [87].
In analysis of the 16S rRNA, we modified the search strat-
Second, DnaSP version 4.0 [88] was used to calculate and egy to overcome the slow rate of convergence and station-
show in graphic form the distributions of observed and arity of the MCMC chains. We analyzed a matrix of 80
expected pairwise nucleotide site differences, also called individuals and 512 bp. We first constructed a UPGMA
mismatch distributions, between all individuals within tree with maximum likelihood distances (model selected
each group, and the respective expected values for grow- in Mr.ModelTest v.2: TrN + I), which was specified as the
ing populations [89]. The model of sudden expansion starting tree in Beast. Previous studies have used values of
describes an initial population at equilibrium, with the 0.33% for the 16S rRNA or 0.7% more generally for a vari-
expected pairwise differences, θ , (θ = 2N µ) and assumes ety of different regions of the mitochondrial genome [e.g.,
0
e
rapid population growth, resulting in θ (Theta final). [78,79]]. We specified the prior for the mean substitution
1
Tau, τ, is the time of the growth measured in units of sub- rate as a normal distribution, with a mean of 0.008 and
Page 15 of 19
(page number not for citation purposes)